Jupyter Spaces is an IPython extension for creating parallel namespaces available within the user namespace. It is designed to be used via IPython magics in Jupyter notebooks.

pip install jupyter_spaces

jupyter_spaces supports Python versions 3.5, 3.6, 3.7 and 3.8.

%load_ext jupyter_spaces

%reload_ext jupyter_spaces

Reloading the extension will remove all spaces.

%%space <space-name>
alpha = 0.50
alpha

When you execute a cell within a space, all references are firstly searched in the space namespace and secondly in the user namespace. All assignments are made in the space namespace.

Trying to delete a user namespace reference will raise an error. Trying to affect a user namespace reference using the keyword global will produce an execution equivalent to not using such keyword.

Mutable objects in the user namespace can be altered (e.g. appending an item to a list).

%remove_space <space-name>

You can access all the spaces' namespaces at once without using any magic. This might be useful to jointly post-process or compare the spaces' contents.

from jupyter_spaces import get_spaces

spaces = get_spaces()
space = spaces[<space-name>]
reference = space.namespace[<reference-name>]

Space objects have two properties:

• name the name of the space
• namespace a dictionary with the namespace of the space

Modifying the spaces via get_spaces will actually modify the underlying spaces.

Many thanks to Yeray Diaz Diaz and Karol Duleba!